| Message ID | 20240115144655.32046-8-pstanner@redhat.com (mailing list archive) |
|---|---|
| State | Superseded |
| Delegated to: | Bjorn Helgaas |
| Headers | show |
| Series | Make PCI's devres API more consistent | expand |
Mon, Jan 15, 2024 at 03:46:17PM +0100, Philipp Stanner kirjoitti: > The bit describing whether the PCI device is currently pinned is stored > in the PCI devres struct. To clean up and simplify the pci-devres API, "PCI devres", 'pci-devres', ... Shouldn't these (and across entire series) be consistent terms? E.g., "PCI devres API". > it's better if this information is stored in the pci_dev struct, because pci_dev struct --> struct pci_dev > it allows for checking that device's pinned-status directly through the > device struct. > This will later permit simplifying pcim_enable_device(). > Move the 'pinned' boolean bit to struct pci_dev. ... > u8 pm_cap; /* PM capability offset */ > unsigned int enabled:1; /* Whether this dev is enabled */ > + unsigned int pinned:1; /* Whether this dev is pinned */ > unsigned int imm_ready:1; /* Supports Immediate Readiness */ > unsigned int pme_support:5; /* Bitmask of states from which PME# > can be generated */ First of all, I think it's better to group PM stuff, like u8 pm_cap; /* PM capability offset */ unsigned int pme_support:5; /* Bitmask of states from which PME# can be generated */ unsigned int imm_ready:1; /* Supports Immediate Readiness */ unsigned int enabled:1; /* Whether this dev is enabled */ unsigned int pinned:1; /* Whether this dev is pinned */ Second, does this layout anyhow related to the HW layout? (For example, PME bits and their location in some HW register vs. these bitfields) If so, but not sure, it might be good to preserve (to some extent) the order.
On Tue, 2024-01-16 at 23:34 +0200, andy.shevchenko@gmail.com wrote: > Mon, Jan 15, 2024 at 03:46:17PM +0100, Philipp Stanner kirjoitti: > > The bit describing whether the PCI device is currently pinned is > > stored > > in the PCI devres struct. To clean up and simplify the pci-devres > > API, > > "PCI devres", 'pci-devres', ... > Shouldn't these (and across entire series) be consistent terms? > E.g., "PCI devres API". Yes, we should agree on a canonical term probably. Probably "PCI devres" is good. > > > it's better if this information is stored in the pci_dev struct, > > because > > pci_dev struct --> struct pci_dev > > > it allows for checking that device's pinned-status directly through > > the > > device struct. > > This will later permit simplifying pcim_enable_device(). > > > Move the 'pinned' boolean bit to struct pci_dev. > > ... > > > u8 pm_cap; /* PM capability offset */ > > unsigned int enabled:1; /* Whether this dev is > > enabled */ > > + unsigned int pinned:1; /* Whether this dev is > > pinned */ > > unsigned int imm_ready:1; /* Supports Immediate > > Readiness */ > > unsigned int pme_support:5; /* Bitmask of states from > > which PME# > > can be generated */ > > First of all, I think it's better to group PM stuff, like ACK > > u8 pm_cap; /* PM capability offset */ > unsigned int pme_support:5; /* Bitmask of states from > which PME# > can be generated */ > unsigned int imm_ready:1; /* Supports Immediate > Readiness */ > unsigned int enabled:1; /* Whether this dev is > enabled */ > unsigned int pinned:1; /* Whether this dev is pinned > */ > > Second, does this layout anyhow related to the HW layout? (For > example, > PME bits and their location in some HW register vs. these bitfields) > If so, but not sure, it might be good to preserve (to some extent) > the > order. As far as I know, one of the greatest weaknesses of C is that neither Ritchie nor the standard ever said anything about the fields' order. Hence, every field-order is purely implementation-defined and in no way portable. So I can't imagine a bitfield will ever directly mapp to a HW-layout? The fields order is purely aesthetic / for the human reader. P. >
diff --git a/drivers/pci/devres.c b/drivers/pci/devres.c index bf957f0bc5ac..de8cf6f87dd7 100644 --- a/drivers/pci/devres.c +++ b/drivers/pci/devres.c @@ -411,7 +411,7 @@ static void pcim_release(struct device *gendev, void *res) if (this->restore_intx) pci_intx(dev, this->orig_intx); - if (!this->pinned) + if (!dev->pinned) pci_disable_device(dev); } @@ -469,18 +469,12 @@ EXPORT_SYMBOL(pcim_enable_device); * pcim_pin_device - Pin managed PCI device * @pdev: PCI device to pin * - * Pin managed PCI device @pdev. Pinned device won't be disabled on - * driver detach. @pdev must have been enabled with - * pcim_enable_device(). + * Pin managed PCI device @pdev. Pinned device won't be disabled on driver + * detach. @pdev must have been enabled with pcim_enable_device(). */ void pcim_pin_device(struct pci_dev *pdev) { - struct pci_devres *dr; - - dr = find_pci_dr(pdev); - WARN_ON(!dr || !pdev->enabled); - if (dr) - dr->pinned = 1; + pdev->pinned = true; } EXPORT_SYMBOL(pcim_pin_device); diff --git a/drivers/pci/pci.h b/drivers/pci/pci.h index dbb76a3fb0e4..3d9908a69ebf 100644 --- a/drivers/pci/pci.h +++ b/drivers/pci/pci.h @@ -809,7 +809,6 @@ static inline pci_power_t mid_pci_get_power_state(struct pci_dev *pdev) * when a device is enabled using managed PCI device enable interface. */ struct pci_devres { - unsigned int pinned:1; unsigned int orig_intx:1; unsigned int restore_intx:1; unsigned int mwi:1; diff --git a/include/linux/pci.h b/include/linux/pci.h index a356bdcc14cc..2f6f44991003 100644 --- a/include/linux/pci.h +++ b/include/linux/pci.h @@ -368,6 +368,7 @@ struct pci_dev { functional, and D3 being off. */ u8 pm_cap; /* PM capability offset */ unsigned int enabled:1; /* Whether this dev is enabled */ + unsigned int pinned:1; /* Whether this dev is pinned */ unsigned int imm_ready:1; /* Supports Immediate Readiness */ unsigned int pme_support:5; /* Bitmask of states from which PME# can be generated */
The bit describing whether the PCI device is currently pinned is stored in the PCI devres struct. To clean up and simplify the pci-devres API, it's better if this information is stored in the pci_dev struct, because it allows for checking that device's pinned-status directly through the device struct. This will later permit simplifying pcim_enable_device(). Move the 'pinned' boolean bit to struct pci_dev. Signed-off-by: Philipp Stanner <pstanner@redhat.com> --- drivers/pci/devres.c | 14 ++++---------- drivers/pci/pci.h | 1 - include/linux/pci.h | 1 + 3 files changed, 5 insertions(+), 11 deletions(-)